INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĪĴ

-0.75

rees

-0.73

aire

-0.72

eus

-0.72

phis

-0.71

ades

-0.70

ade

-0.70

pter

-0.70

INAL

-0.68

ais

-0.67

POSITIVE LOGITS

gger

0.73

 cryst

0.70

Eva

0.67

quin

0.64

jac

0.60

oplan

0.60

hed

0.59

beh

0.59

morph

0.58

 Mord

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.